Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.



When should I stop the simulation?
Posted:
Jun 16, 2012 6:54 AM


I do a simulation to calculate a parameter of an experiment.
I generate N normally distributed random numbers and then I calculate the parameter from these numbers (it's a correlation coefficient and in my case its value is always big: 0.98...0.9999).
As the simulation goes on, I periodically print the parameter and I see that it converges to the real value (the left column is the number of the samples I use to calculate the parameter):
262144) 0.9988634 524288) 0.9988620 ... 2359296) 0.9988636 2621440) 0.9988639 2883584) 0.9988641 3145728) 0.9988642 3407872) 0.9988642 3670016) 0.9988641
Is there any method to decide when to stop the simulation?
Thanks Cristiano



